© KROHNE 07/2003

GR

7.02337.22.00

Fundamentals of

Radar Technology
for Level Gauging
4th edition (revised and expanded) by Dr.-Ing. Detlef Brumbi KROHNE Messtechnik GmbH & Co. KG

1
Foreword to the 4th edition

The first edition of "Fundamentals of Radar Technology for Level Gauging" in 1995 originated from the wish to write a concise account of the technical basis of the then still relatively young field of industrial radar technology. Although countless books and articles had already been published on radio frequency technology and radar methods, there was still a general lack of information on the special issues relating to level measurement. Owing to popular demand, this booklet now appears in its 4th edition. Following the additions made to the 3rd edition, the contents of this new edition have again been updated, adapted to the latest state of the art and greatly expanded. Thus many aspects have been annotated by additional information, formulae of calculation and diagrams, and more space given to the subjects of TDR and signal evaluation. This booklet is not a brochure for a particular industrial product but a well-grounded article on technical fundamentals to explain the processes taking place in radar and TDR level gauges. For some it may serve as a kind of textbook to provide a deeper understanding of level measurement technology, for others it may serve as a reference work to provide more details on specific matters. Not included are subjects of a general nature, such as bus systems and power supply concepts (e.g. 2-wire technology), since they are generally applicable to process measurement technology and are dealt with extensively in many works.

Duisburg, January 2003 Detlef Brumbi

Radar handbook

Content 1 Introduction 1.1 RADAR systems 1.2 Radar milestones 2 General 2.1 Frequency, wavelength and propagation rate 2.2 Electromagnetic frequency spectrum 2.3 Postal regulations 2.4 Hazards from microwaves 2.5 Fields of application 3 Radar level measurement systems 3.1 Overview of level measurement methods 3.2 Radar level measurement - General 3.3 Comparison of radar methods 3.4 Interferometer radar 3.5 Pulse radar 3.5.1 Bandwidth of an HF pulse 3.6 FMCW radar 3.6.1 Principle 3.6.2 Type model 3.6.3 Frequency control by PLL 3.7 Power balance (“radar equation”) 4 Components for radar systems 4.1 Active devices 4.1.1 GaAs transistors 4.1.2 Active diodes 4.1.3 Silicon devices 4.1.4 Velocity-modulated tubes 4.2 Oscillators to generate microwave frequency oscillations 4.2.1 Fixed-frequency transmitters 4.2.2 DRO 4.2.3 VCO 4.3 Circuit stages for processing radar signals 4.3.1 Mixers 4.3.2 Receiver noise 4.4 Line transmission 4.4.1 Coaxial cable 4.4.2 Twin line 4.4.3 Planar lines 4.4.4 Wire in free space 4.4.5 Waveguide 4.4.6 Coupling into waveguides 4.4.7 Directional coupler 4.4.8 Reflections at junctions 4.4.9 Plug connectors 5 Antennas 5.1 Types of antenna 5.2 Antenna gain 5.3 Radiation angle 5.4 Polarization 5.5 Directional diagrams
2

page 5 5 5 6 6 6 8 8 8 9 9 10 10 11 12 12 13 13 13 14 15 16 16 16 16 17 17 17 17 17 17 18 18 18 19 19 19 20 20 20 21 22 22 23 24 24 24 25 25 26
Radar handbook

7.6 Signal evaluation in the pulse-radar system 8.2 Transmission through dielectric windows 6.2 Signal-to-noise ratio 8.3 Dielectric permittivity 7.4 Liquid mixes 7.1 Reflection factor 7.Content 6 Wave propagation 6.4 Atmospheric signal attenuation 6.2 Tank bottom tracing 8.1.3.8.4 Scattering from particulate materials 7.8.6 Angle of reflection 8 Evaluation methods 8.2 Propagation rate on lines 6.3.1 Useful signal 8.2 Reflection at interfaces 7.1 Chemico-physical assessment 7.5 Reflected radar cross-section from limited targets 7.3 Propagation rate in stillwells/waveguides 6.3.7.3.3 Temperature and viscosity dependence 7.1 Influence of the carrier medium (atmosphere) 6.5.3 Phasegradient method 8.5 Modified radar equation 6.7.4 Interference 8.7 Signal evaluation in FMCW radar 8.7 Propagation along electric lines (TDR method) 7 Reflection 7.3 Free-space attenuation 6.1.5 Bulk (particulate) materials 7.1 Counting method 8.2 Unambiguity 8.4 Tracking 8.8 Special methods 8.5.5 Sample calculation for received powers 8.3.3 Measurement uncertainty 8.3 Signal-to-interference ratio of an interference reflector 8.1 Empty-tank spectrum 8.2 Fourier transform 8.3 Interface detection Annex A Table of dielectric permittivity B System-theoretical comparison between interferometer and FMCW methods C Bibliography page 28 28 28 30 31 31 32 32 34 36 37 38 38 39 40 40 42 42 42 43 43 44 44 45 45 45 46 46 48 48 48 48 49 50 50 50 51 52 54 54 55 56 58 58 60 63 Radar handbook 3 .1 Propagation rate 6.7.6 Equivalent isotropic radiation power (EIRP) 6.8.5.2 Frequency dependence 7.1.1 Object discrimination 8.

power density Normal pressure P PE PS r R RS t T TN v ZL α εo εr εr’ εr’’ εr. spectral line interval Fundamental frequency Cut-off frequency in the waveguide Sampling frequency Cut-off frequency Doppler frequency Repetition frequency Sweep frequency. noise figure Linearity error Antenna gain Tank height Shifted tank height Magnetic field strength Boltzmann constant (1. delay Temperature (in Kelvin). velocity of light Speed of light in vacuum (~3·108 m/s) Conductor diameter Diameter (of waveguide.Symbols used a ∆a A AE AR b B c c0 d D D1. antenna) Propagation loss Free-space attenuation Electric field strength r.s. number of sampling points Pressure.D2 Dfree E Eeff Ep EIRP f ∆f f0 fc fA fg fD fi F ∆F G1. G2 h hv H k K L N p pN Distance. spacing Object discrimination.38 ·10 –23 J/K) Correction factor Filling level Integer. η2 σ τ ϑ 4 Radar handbook . sweep time Normal temperature Propagation rate in the medium. transit time. value of electric field strength Peak value of electric field strength Equivalent Isotropic Radiation Power Frequency Frequency difference. speed of target Natural impedance Propagation loss factor Absolute dielectric permittivity in a vacuum (8. measuring error Aperture area Receive area Reflection area Width Bandwidth Speed of propagation.N Power Received power Transmission power Voltage reflection factor (Power) reflection factor Reflection scattering from a particulate material Time.854 ·10 –12 As/Vm) Relative permittivity Real part of relative permittivity Imaginary part of relative permittivity Relative permittivity of gas at normal conditions Phase Phase difference Wavelength Cut-off wavelength in the waveguide Wavelength in free space Radiation (antenna) efficiency Radar cross-section Pulse duration Temperature (in °C or °F) ϕ ∆ϕ λ λc λ0 η1.m.

but often just one is used for both transmitting and receiving the radar signal. civil and industrial applications — has developed rapidly. 1: Basic structure of a radar system Receiver Antenna Transmission path 1. as outlined below: 1865 1887 1904 1922 1935 from 1939 c. Transmitter Antenna Transmission path Reflecting Target Fig. GB) Intensive research for military applications (GB. a transmission path.2 Radar milestones Even though the existence of electromagnetic waves had been predicted by Maxwell in the 19th century and the theoretical principles laid down. D) Radar devices to monitor the speed of road traffic First radar level gauge First compact radar level gauge Radar handbook 5 . Introduction 1. a further transmission path (usually identical with the first one). when for the first time it was possible using a continuous-wave radar with 5 m wavelength to detect a wooden-hulled ship. the reflecting target. USA. and a receiver with antenna. The term RADAR is an acronym from RAdio Detection And Ranging A complete radar measuring system is comprised of a transmitter with antenna. Two separate antennas may be used. USA) Used for locating aircraft (Watson-Watt. 1960 1976 1989 Theoretical prediction of electromagnetic waves (Maxwell) Experimental verification of Maxwell’s theory (Hertz) Patent: “method of signalling distant metallic objects to an observer by means of electric waves” (Hülsmeyer) First radar device (Taylor & Young. Since then radar technology — for military. the technical means for constructing a radar device was not available until 1922.1 1.1 RADAR systems The term “radar” is generally understood to mean a method by means of which short electromagnetic waves are used to detect distant objects and determine their location and movement.

in a vacuum it amounts to exactly 2 c0 = 299 792 458 m/s. The unit "metre" is now derived from c0 (previously vice versa). 2. H (6-8 GHz). microwave frequencies are used intensively for communications and locating purposes. 1 2 3 The power or power (flux) density (power/area) is normally used as the measure of intensity. wavelength and propagation rate To characterize electromagnetic waves. C. visible light and ultraviolet ranges. K (20-40 GHz). G (4-6 GHz). The following interrelationships exist: c=λ·f λ= c f f= c λ The propagation rate c is equal to the velocity of light. D. bands upwards of 100 MHz with consecutive coding A. whose commonly used letter code is also given in Figure 2. E. the relevant factors in addition to intensity1 are their frequency f and wavelength λ. microwave frequencies are used up to approx. 120 GHz – a limit that will extend upwards as technology advances. The speed of light c0 is now defined as an absolute natural constant (since 1983).2 2. The 4 to 120 GHz frequency range is divided into 7 bands3. Radar handbook 6 . However. or the electric or magnetic field strength.2 Electromagnetic frequency spectrum Microwaves are generally understood to be electromagnetic waves with frequencies above 2 GHz and wavelengths of less than 15 cm (6”). or approx.1 Frequency. F. B. Far above this limit are to be found the infrared. in gases it is only negligibly lower. For technical purposes. General 2. J (10-20 GHz). there are different divisions and codings. that are linked by way of their propagation rate c. I (8-10 GHz).g. As shown in Figure 2 on the next page. e. 3·108 m/s.

2: Electromagnetic frequency spectrum with typical applications in the microwave range Radar handbook 7 . Radar-Füllstandsmesssysteme Fig.2 3.

for example.8 GHz ± 50 MHz ± 75 MHz 24. altimeters ● speed measurement in road traffic ● distance warning for vehicles ● meteorology ● materials analysis. ● industrial level measurement systems 4 5 6 For example. 2.5 Fields of application Microwave techniques and radar systems have become established in many technical areas – for military.125 GHz 61. Some of these are shown in Figure 2. Most countries require an approval or licence from the postal or other authorities. limiting of the transmission power..7 mW/cm2 [BG] and 5 mW/cm2 (ANSI). however. These define e. Taking.2 3. scientific and medical purposes (so-called ISM bands). Radar-Füllstandsmesssysteme 2. notification of the location. while the European Union requires these to conform to the R&TTE (Radio and Telecommunications Terminal Equipment) directives. according to the present state of knowledge it can be assumed that persons are not at risk provided the 3 directives issued under DIN-VDE 0848 [DIN] and by the “Berufsgenossenschaft der Feinmechanik und Elektrotechnik” [BG] (German employers’ liability insurance association in light engineering and electrical engineering) and ANSI are observed. for navigation. Even at this critical point this is far below the specified limit values.45 GHz 5. 10 mW from microwave level measuring systems.4 Hazards from microwaves The health hazard potential of electromagnetic waves is a highly controversial subject. civil and industrial purposes.3 Postal regulations To prevent mutual influence and interference.. the use of microwaves is officially regulated. the maximum power density at the aperture6 of an antenna with 100 mm (4”) diameter is 130 µW/cm2 (20 µW/in2). restriction to certain local areas. provision of shielding equipment. but the requirements are far less strict. However. chemical analysis ● humidity recorders. Currently these are the following 4 frequency ranges: 2.25 GHz ± 125 MHz ± 250 MHz 2. The following brief overview demonstrates part of the large range of application: ● to locate and measure the movement of flying objects ● to locate aircraft and ships.g. limits for the power density: 1 mW/cm2 [DIN] and 6. A post office licence often involves compliance with special conditions4. There are.1 . typical transmission power levels of 0. Homogeneous power distribution assumed Radar handbook 8 . also internationally released5 frequency bands for industrial. An approval or a licence is normally required.

a) The position of a float. dependent on liquid. good accuracy. whose density must be less than that of the liquid. A term frequently used is “bottom pressure transmitter”. temperaturedependent resistor is dipped into a liquid. Normally only used as a liquid-level switch.3 3. good accuracy. Measures the current flowing through an electrode at the instant it comes into contact with the liquid. travel along a conductor dipped into the liquid. Advantages: relatively low cost. b) a float (“displacer”) dips partially into the liquid and the change in weight is measured. expensive. low accuracy. of which the most important are described in brief: Hydrostatic: A pressure sensor is fitted to the tank bottom to measure the differential pressure relative to the environment. c) a sensing plate is mechanically guided on the surface until uplift is detected. Drawbacks: risk of contamination. is sensed. almost independent of carrier medium and surface of the liquid product. Given an electrically conductive liquid the electrode must be insulated. Deduces the transit time of a laser beam reflected from the surface of the liquid. narrow beam angle. An ultrasonic signal is emitted. Measures the level-dependent capacitance between an electrode dipped into the liquid and the tank wall. 9 Buoyancy: Capacitive: Conductive: Vibration: Thermal: Radiometric: Laser: Ultrasonic: Microwave: TDR: Radar handbook . Advantages: non-contact measurement. Measures the transit time of the signal. even very weak reflections detectable. Drawbacks: dependent on the medium. Advantage: non-contact measurement.1 Overview of level measurement methods Measuring the level of liquids or solids in vessels is a frequent requirement in industry. Radar level measurement systems 3. Also called "directed microwave". in which case the insulating capacitance in the wetted part is effective. Drawbacks: dependent on the density of the medium. Advantage: very simple. elaborate calibration. low accuracy. Gamma rays are more greatly attenuated as they pass through the medium than in the atmosphere. Drawbacks: radiation exposure. fails in vapour atmospheres. however. Advantages: largely independent of tank internals. fails in vacuum conditions and vapour atmosphere. Utilizes the greater heat dissipation when a current-carrying. Advantages: non-contact measurement. Drawbacks: low accuracy. reflected from the surface of the liquid and received. Drawbacks: sound velocity heavily dependent on gas composition and temperature of the atmosphere. Advantage: low cost. A method that also measures the transit time of radio-frequency signals which. heavy contamination causes failure. Many traditional and modern methods have been developed [Webster]. the electrical resistance varies with the depth of immersion. Measures the transit time of a radar signal that is reflected from the surface of the liquid. Advantage: noncontact measurement. very good accuracy. Drawbacks: dependent on the density of the medium. Measures the degree of damping of a vibrating fork when dipped into the liquid. Normally only used as a liquid-level switch.

Radar-Füllstandsmesssysteme 3. 6. for which they require approx. the waves travel a distance of 2 m. This is the method used for vehicle speed checks. the speed can be calculated from the Doppler shift of the frequency.2 Radar level measurement .3 3. Radar handbook ● ● ● 10 . The distance of the reflecting boundary layer – independent of the radar method used – is determined by way of the transit time t of the microwave signal: per metre target distance. 3: Geometric configuration of reflectors and signal strength as a function of distance useful signal 3. The velocity v of moving targets can be determined by the Doppler shift in the received signal. FMCW radar (frequency modulated continuous wave) the signal is continuously present but the frequency is modulated. At the same time. The absolute distance information is. reflected from the surface of the product and the echo received again after a time interval t. λ/2-periodical (see chapter 3. 2 ns for 1 ft distance). Pulse radar: transmits a radar signal in short-duration pulses (carrier-modulated or nonmodulated). Generally the distance measured is a = c · t / 2.6).7 ns (or approx. signal strength distance interference signal Fig. usually in successive (linear) ramps. however.3 Comparison of radar methods The following gives a brief description of the signal forms and properties of commonly used radar methods – keyed to specific application requirements: ● CW radar (continuous wave): this system transmits a continuous signal of constant frequency f.4). Interferometer radar: the phase of the received signal relative to the transmission phase can be established in order to measure changes in distance with the aid of a non-modulated high frequency signal of constant frequency. The level is then calculated from the difference between tank height and distance. The Doppler frequency is: fD = 2 · v · f /c. Distances cannot be measured. The distance of the target is deduced from the transit times of the pulses from the transmitter via the reflecting target back to the receiver (see Section 3.5).General A radar signal is emitted via an antenna. The distance of the target can be deduced from the received signal (see Section 3. however.

for example. TDR method (time domain reflectometry): this is similar to pulse radar. Hence.g. This signal is reflected from a reflector (e. ● ● The basic methods used for radar level measuring equipment are pulse radar or FMCW radar. and the resultant phase difference ∆ϕ determined from the received signal: forward wave λ reflected wave ∆ϕ reflector Fig.4 Interferometer radar With this method.3 3. In another method. 3. e. the characteristics of absorber materials or the moisture content of products. pulses are frequency-modulated („chirp“ radar). Radar-Füllstandsmesssysteme ● Reflectometer radar: this method is used to measure the complex reflection coefficient of the target. Radar handbook 11 . 4: Principle of interferometer radar For this purpose. Combined methods: a combination of reflectometer and pulse can. sometimes supported by the interferometer method. these processes are described in greater detail in the following sections. but is normally conductor-bound and used with electrical pulses without carrier frequency. also measure absolute distances. a microwave signal of constant frequency is transmitted for a certain period of time. liquid surface). From this material information can be deduced. But the result is periodical with N · λ/2 and therefore ambiguous. phase evaluation is carried out between the transmitted signal and the received signal with a delay of t = 2a/c: The accuracy of the interferometer method is determined by the resolution of the phase measurement and can be very high.g.

Radar handbook 12 .003 µs. Fig.5 Pulse radar The principle is very simple: a short electrical pulse or wave package is transmitted. 6: Time signal and spectrum of a radar pulse The same diagram applies to an (ideal) squarewave pulse without carrier frequency when f0 = 0. the requirement for a time accuracy of ps for the sampling procedure remains the same. 3. an accuracy of approx. for example. the signal bandwidth can be employed as a reference quantity: a pulse radar with 1 ns pulse duration has.g.1 Bandwidth of an HF pulse Basically a frequency spectrum can be assigned to every time signal (Fourier transform). the same bandwidth as an FMCW radar with 1 GHz sweep. time signal (HF pulse): frequency spectrum (Fourier transform): sinc Fig. a spectrum is formed representing a sinc function (sin(x)/x). for a signal with 1 MHz repetition frequency at the points: 0. The 3 dB bandwidth is approx.6).5. see Fig. Nevertheless.04”) distance measurement.004 µs etc. B = 1/τ. An extended time scale is normally used by sampling7 so that signal evaluation can be performed in the range of low frequencies (see chapter 8.000 µs.002 µs.6). 1. If a high-frequency signal of constant frequency f0 is pulsed with duration τ. meets the reflector after time t1 = a/c and is received back after a total time t2 = 2a/c.001 µs. 6). 5: Pulse radar measuring system 3. 2. 7 Sequential sampling of a periodic signal at continuously delayed points of time. e. Radar-Füllstandsmesssysteme 3. 4. 6 ps time measurement is needed. The technical difficulty lies in obtaining accurate measurement of the time t2 because for an accuracy of 1 mm (0.3 3. Another requirement is a good repeatability of the reflection signals during a sampling sequence. To be able to compare pulse radar systems with FMCW systems (see chapter 3.

8 shows an example of an FMCW radar system 9.6 FMCW radar 3. 13 Radar handbook . The frequency f of that signal is proportional to the reflector distance a. 8 9 As opposed to the pulse radar method The individual electronic components are described in Section 4. the frequency of the low-frequency mixed signal remains constant during the sweep procedure.2 Type model Fig. The received signal is decoupled via a directional coupler. in this method. 7: Functional principle and signal trend in FMCW radar Due to the time delay during signal propagation.6. the differential frequency is formed by mixing. the transmitted frequency changes so that from the difference between the momentary transmitted frequency and the received frequency a low-frequency signal (typically up to a few kHz) is obtained. mixed with the transmission signal. and processed by the microprocessor.3 3. therefore. 3. The instantaneous frequency needs to be measured in order to ensure good sweep linearity. If the frequency sweep is linear. Radar-Füllstandsmesssysteme 3. Because the resultant signal frequencies8 are low. linearly in a time interval (frequency sweep). transmitter differential frequency f antenna delay time t = 2a/c frequency a transmitter reflector time receiver mixer f Fig. Normally evaluation is by means of digital signal processing. A variable oscillator VCO is controlled by a microprocessor so that the desired frequency sweep is obtained. This is done by counting the frequency after it has been mixed with a known frequency (DRO). further signal processing is technically simple and very accurate.1 Principle FMCW radar uses a linear frequency-modulated high-frequency signal.6.g. the delay t is transformed into a frequency (df/dt is the sweep velocity): f = df/dt · t Technically. This signal is amplified and fed via a pin coupler into the transmitting antenna. the transmission frequency rises e.

which dynamically sets the required frequency precisely at every instant of the sweep. Radar-Füllstandsmesssysteme directional coupler received signal antenna VCO amplifier pin coupler DRO Fig.2]: ∆a/a < 8·∆F/F To obtain measuring accuracy in the mm range at distances of 10 m and more. but with the added benefit of gaining definitive information on distance. the same measuring accuracy can be obtained as with a static interferometer radar system (see Annex B). 8: Basic circuit diagram of an FMCW microwave circuit mixer transmitted frequency mixer microprocessorcontrol distance measuring signal 3.3 3.6. 14 Radar handbookl . nonlinearity of the frequency must be in the order of 10 –6. This can only be accomplished by using active frequency control by means of a PLL circuit (phase locked loop) [Musch].3 Frequency control by PLL The measuring accuracy of an FMCW system depends on the non-linearity of the frequency sweep [Stolle. Microprocessorcontrol Referenceoscillator Microwave oscillator Phasedetector Fig. 9: Structure of a PLL frequency control Loop filter Antenna With such a system.

7 Power balance (“radar equation”) With reference to the basic system shown in Fig. only emit as much power as is fed into it. 15 Radar handbook . The antenna gain describes the higher power density that is obtained with a directional antenna. when compared with an isotropic antenna. of course. See also Section 5. the following system equation – often termed “radar equation” – is obtained: PE = PS · G1 · R · G2 D1 · D20000 10 An antenna can. D2 reflection factor R antenna gain G2 received power PE Accordingly.3 3. 1 the following power balance quantities are assigned to the various component parts: Transmitter: Transmitting antenna: Transmission paths: Reflecting object: Receiving antenna: Receiver: transmission power PS antenna gain10 G1 propagation loss D1.2. Radar-Füllstandsmesssysteme 3.

Such active diodes include: ● the tunnel diode. as shown schematically in Fig.1. More recent developments are hetero-bipolar transistors (HBT) and high electron mobility transistors (HEMT).1 GaAs transistors Transistors made of gallium arsenide are amplifying semiconductor devices (usually in the form of MESFETs. Components for radar systems 4.1 Active devices 4. with very few components and a very small space requirement.e. would only be implementable with considerable outlay. Radar handbook 16 . For example. since here the GaAs MESFETs have currently reached the limit of their technical capabilities. 11 Strictly speaking. GaAs or GaSb) ● the IMPATT diode (impact avalanche transit time) ● the BARRITT diode (barrier injection transit time) 11 ● the transferred electron device (TED) or Gunn diode ● the quantum effect diode. but it is nevertheless ranked among the active diodes because of similar physical transit time effects. as required for FMCW systems. HEMT (b) and HBT (c) (after [Meinke]) n+ n n+ n+ GaAs GaAs Source Gate n AlGaAs semiconductor substrate (undoped GaAs) (b) HEMT Drain Emitter Basis AlGaAs n GaAs p GaAs np+ Collector semiconductor substrate (undoped GaAs) (a) MESFET n+ GaAs semiconductor subtrate (c) HBT 4. Gunn diodes are therefore used mainly for pulse oscillators in the range above 20 GHz. For example. that are used principally in the frequency ranges of 1 GHz to approx. Source Gate Drain Fig. this device is not a diode because it has no p-n junction. these devices are used in satellite receivers and other high-frequency circuitries for oscillators. also known as the Esaki diode (made of Ge.g. e.2 Active diodes Various bipolar semiconductor devices are available which have a negative differential resistance characteristic in certain operating conditions and so can be used for amplifier or oscillator purposes. Frequency modulation. 30 GHz for low-power applications (a few mW). 10 Cross-section through the structures of MESFET (a). they can be used to assemble low-power oscillators in the microwave range up to about 100 GHz. field-effect transistors with metallic gate). mixers and amplifiers. These types of transistor consist of various semiconducting layers.4 4. 10. resonant tunnel diode (RTD) Gunn diodes are relatively expensive and therefore not worthwhile for low frequencies < 20 GHz.1. i.

which e.1.1. 4. Considerable advances have been made in the last few years. two metallization layers and a ”buried" emitter area.g. 13 The transit frequency f T is the frequency extrapolated from the decreasing frequency characteristic by –20 dB/decade.4 4.2 DRO An oscillator for a stable fixed frequency can be made with the aid of a dielectric resonator (usually made of ceramic material) in conjunction with an amplifier element (e. 4.1 Fixed-frequency transmitters Oscillators for systems with a constant transmission frequency (CW radar) are usually equipped with GaAs-FET. 4. 11: Schematic of an oscillator circuit with transistor Tr and varactor diode Cvar 12 Unlike the structure of a standard n-p-n or p-n-p bipolar transistor.2.3 VCO A voltage controlled oscillator is required for transmitters transmitting frequency-modulated signals.g. SIEGET or Gunn diodes in known basic circuitries. a DRO is frequently used as a reference for the mixer to determine the transmission frequency (see Fig.3 Silicon devices The development of transistors made of silicon (Si) is continuing in the direction of higher cut-off frequencies. Utuning resonant circuit Fig.g. magnetron.45 GHz). with the SIEGET family12 (Siemens grounded emitter transistor).2. which currently is available with transit frequencies13 of up to 45 GHz and achieves sufficient gain in the range up to about 15 GHz. for example. Within this decreasing range the gain is approx.: G = fT/f.g. an extremely stable frequency is guaranteed with low temperature drift. acts on a variable capacitance diode (varactor diode) in a resonant circuit. Radar handbook 17 . 8). In measuring systems.2. the transmitted frequency can be varied – which is necessary for an FMCW radar system (see Figs 8 and 11). Velocity-modulated tubes are not suitable for radar level measuring systems because they are too large in size and such high power outputs are not required. By means of a control voltage. GaAsMESFET. at which the gain is 1 (0 dB). klystron). SIEGET). Special velocity-modulated tubes are available to generate frequencies up to some 100 GHz.4 Velocity-modulated tubes High microwave outputs (up to some kW) at high efficiency levels (80%) can be obtained with velocity-modulated tubes (e. when it is termed a DRO = dielectric resonance oscillator. Magnetrons are used. 4. in microwave ovens (2. e. 4. Since the resonant frequency depends essentially on the geometric dimensions of the resonator.2 Oscillators to generate microwave frequency oscillations 4. the frequency-limiting parasitic effects are minimized in the SIEGET by short bonds.

8) generally feature two mixers: ● one to allow measurement of the VCO transmitted frequency after mixing with the DRO frequency (e. the input-related noise is increased by the noise figure F : P’noise = F · k · T · B.5). with the aid of transistors in various circuit configurations (as multipliers or non-linear amplifiers) or with diodes using their non-linear characteristic. 18 Radar handbook .g. the differential frequency is processed as a distance-proportional signal for level measurement. FMCW systems (see Fig. For a receiver. The signal-to-noise ratio should be as high as possible in order to obtain high detection reliability and a low error rate. VCO = 10 GHz and DRO = 9 GHz.3.1 Mixers Mixers are used to generate an output signal from two signals of different frequencies with the appropriate differential frequency. where k is the Boltzmann constant. another to mix the signal received by the antenna with the transmission signal.3 Circuit stages for processing radar signals 4. The latter is generally eliminated by frequency filtering. which is easier to process metrologically than the considerably higher VCO frequency by the direct method). Multiplying two sine functions together produces sinusoidal signals with the differential and the cumulative frequency. and B the receiving bandwidth. powers of less than 1 mW to a few mW are sufficient to obtain a sufficiently large signal-to-noise ratio. taking into account the total transfer function (see Section 6. Mixers can be made e.g. The required transmission power is determined from this. With the relatively short ranges (up to some 10 s of metres or 100s of feet) that are relevant for level measurements. ● 4. giving a mixture frequency of 1 GHz.2 Receiver noise Natural thermal noise is calculated according to: Pnoise = k · T · B. T the absolute temperature. Radar-Füllstandsmesssysteme 4.4 3.3.

The electromagnetic field surrounds the entire space around the twin line.35 with a Teflon filling (εr = 2. 12: Coaxial cable The natural impedance ZL is an important parameter of a transmission line. It is suitable for signals from d.c.2 Twin line The twin line consists of two conductors routed in parallel by spacers or a dielectric jacket. The electromagnetic field is only formed inside the cable.1 Coaxial cable A coaxial cable generally consists of a wire as the internal conductor and an external conductor – wire mesh or tube – with synthetic material in between as the dielectric.4. Depending on quality. to a few GHz.3 with an air filling and D/d = 3. outer conductor dielectric inner conductor Fig.4 Line transmission 4. so the coaxial cable is a low-radiation type. It describes the ratio between voltages and currents in the individual waves on the line.4 4. 4.1). 20 GHz. 13: Twin line Natural impedance is calculated according to the equation: 14 Assuming a lossless line and non-magnetic environment (µr = 1) 19 Radar handbook . dielectric 2 wires Fig.4. the diameter ratio must be D/d = 2. and is calculated for the coaxial cable according to the equation14: To design a "standard" line with 50 Ω. coaxial cables are capable of transmitting electrical signals from direct current up to high-frequency waves of approx.

14. 7 GHz at D = 25 mm (1”).7) with a rod or wire rope.3 Planar lines Planar lines (also called striplines or microstrip lines) consist of plane line structures applied to a dielectric substrate. the waveguide can only transmit signals with a defined minimum frequency fc. the line width must be b = 0. 14: Planar line rear metallization In the structure shown in Fig.4. This is roughly the same setup as that of a TDR system (see chapters 3.5 Waveguide A waveguide consists of a metal tube with circular or rectangular cross-section in which high-frequency electromagnetic waves are propagated along its length. 14.75 mm if the natural impedance is to be 50 Ω. The advantage of this form of line is that other devices are also easy to mount. The waveguide can be filled with air or a dielectric. In contrast to the coaxial cable or the twin line.4.4.4. for example. 4.1]: For example. waveguide dielectric Fig. 4. The natural impedance ZL = 377 Ω /√εr is approximately the same as that of free space (so-called characteristic impedance).4 4.2) of thickness a = 0. for a commonly used Teflon substrate (εr = 2. An example is shown in Fig. For the H11 fundamental wave in the circular waveguide with an inside diameter D.1 and 6.25 mm. The following formula applies: 20 Radar handbook . fc is approx. Wire in free space A single wire in free space also transmits electromagnetic waves – however with losses because of radiant emittance. for b ≥ a the natural impedance is approximately [Pehl.

since then higher modes can also be propagated at a different velocity: which is slower than in free space. 15: Waveguide (left) and coupling into waveguide (right) waveguide In addition.g.6 Coupling into waveguides A coupling device is required for transition from the double-line transmission path (e.820 E21 0. Radar-Füllstandsmesssysteme However. a so-called pin coupler.4 3.612 H12 0. for example: Typ λc /D H11 1.4.306 H21 1. e. The cut-off wave length λc for individual modes in a circular waveguide is.706 E01 1. there are various possibilities for planar coupling – directly from the PCB to the waveguide.589 At GHz frequencies. coaxial cable) to a waveguide. 4.029 E11/H01 0. 16: Planar waveguide coupling with fin line (left) or in radial arrangement (right) lower side waveguide Radar handbook 21 . coaxial connection sheath air pin coupler Fig. waveguide upper side PCB Fig. a situation to be avoided is for the operating frequency to be distinctly higher than the cut-off frequency fc. the transmission losses of a waveguide are lower than those in coaxial cables or twin lines.g.

For single-antenna systems it needs to be fitted in front of the antenna connection in order to decouple the received signal and separate it from the transmission signal. 17: Examples of planar directional couplers [Pehl. The same applies analogously when a wave is fed into another gate. 0). Where there is a junction between one line structure with natural impedance Z1 and another line with Z2 we obtain the following voltage reflection factor r: 22 Radar handbook . the transmission amplifier could be connected to gate 1. Gate 3 is not used and must feature a non-reflecting termination. the antenna to gate 2. 4. the natural impedance will also change.8 Reflections at junctions At every junction in a conductor (particularly in waveguides) at which the cross-sectional geometry or the dielectric material changes. and the receiver mixer to gate 4. For example.7 Directional coupler A directional coupler transforms only such waves along a second line structure that are transmitted in a predetermined direction along the first line.4 4. 17 is that a wave (P1 fed into a gate is transmitted only to 2 gates (P2 and P3) while the 4th gate remains largely without power (P = approx. 8.4.4. P3 3 P≈0 4 4 P≈0 3 P3 Fig. For that reason suitable design measures are necessary at junctions in order to keep reflections as small as possible. since both signals are present simultaneously at the coupling point. By this means the high-energy transmission signal is not transmitted to the sensitive mixer. so that unwanted reflections can occur at these points.1] 1 P1 2 P2 P1 1 2 P2 The characteristic of the planar structures shown in Fig. in the system shown in Fig.

They exhibit low throughput losses and good active return losses. 19). 15 Mostly dielectric bodies with gradually or abruptly changing cross-section or filling of the waveguide. By combining different materials with matched εr and using special geometries15 an approximately reflection-free junction can be obtained which effects “impedance transformation” of the waves (Fig. the challenge is to implement a junction that has the least possible reflections. when the waveguide is filled with materials having different dielectric constants er.9 Plug connectors Plug connectors are required to provide quick-connection facilities between coaxial cables or between such cables and circuit modules. ε0 ε1 ε2 ε0 ε1 ε2 Fig. Fig. So-called SMA connectors are the most commonly used for the maximum-frequency range up to 26 GHz.4 A negative value for r means that the polarity of the reflected pulse reverses. 23 Radar handbook .4. the geometry being in a certain proportion to the wavelength in the dielectric. 18: Voltage reflection factor at the junction of a 50 Ω line In waveguide structures. 19: Examples of impedance transformers in waveguides λ λ 4 4 λ 4. 18 shows the value of r at the junction between a (standard) 50 W line and a line with natural impedance Z2: Fig.

2]. [Philippow. power density in the lobe is higher. In addition to adjusting natural impedance. auxiliary reflector pyramidal horn conical horn primary radiator parabolic reflector dielectric rod antenna coupling point dielectric Fig.1 Types of antenna The following figure shows some of the most commonly used antenna types. [Pehl. as it were. e. the antenna “amplifies”.g.2 Antenna gain The antenna emits the waveguide wave into the free space below the antenna aperture. 20: Various form of antenna dielectric sleeve dipole dielectric lens 5. it also has a directional effect. The antenna gain is the better the larger is the aperture area A of the antenna and the smaller the wavelength λ. The quantity “antenna gain” is closely connected to the directional effect: since the high-frequency power is emitted in a narrower spatial angle.3].5 5. the signal. [Meinke] Radar handbook 24 . The spectrum of geometries is extremely varied16. but the forms shown represent the most practical versions for level measuring systems. Antennas 5. The following relation exists between the antenna diameter D or aperture area A of a conical horn radiator and the antenna gain G1: 16 Refer to the literature.

the larger the aperture area.e.e.0.5 3. For level measuring systems. i. 17 Reciprocity of characteristics for transmission and reception 25 Radar handbook . The smaller the radiation angle. Radar-Füllstandsmesssysteme Typical values for antenna efficiency h1 are approx. 40°. at the edge of this lobe the power density is half the size it is in the middle).3 Radiation angle A characteristic quantity for describing the directional effect is the radiation angle or halfvalue width.5 .2]) The radiation lobe is slightly asymmetrical (elliptical) due to polarization of the waves (see subsection below). a small radiation angle. Fig.e. 21 gives a rough estimation of the half-value width of horn radiators with an aperture angle of approx. By approximation the following equation applies (D = antenna diameter): angle [degrees] 90 80 70 60 50 40 30 20 10 0 0 1 2 3 radiation angle 4 5 6 7 8 D 9 10 λ Fig. good focusing. 21: Radiation angle of horn antennas (approximative values after [Pehl. is desirable in order to avoid interference reflections as much as possible from the tank wall or tank internals. Due to reciprocity17 a gain G2 is imputed to a receiving antenna that is generally equal to the transmission G1: G2 = received power of antenna in a planar wave field (optimally aligned) received power on an ideal isotropic radiator The following correlation with the effective receiving area AE is given by: 5. the higher the antenna gain. This is defined as the cone angle at whose edge the power density is 3 dB below the maximum power density (i. i. 0.8.

represented by different radiation patterns for the E and H field is not discussed here.5 3. Any asymmetry due to polarization.g.5 Directional patterns The radiation pattern of an antenna describes the distribution of power density over the solid angle. From the examples given in Fig. Vectors E and H are always perpendicular to each other and perpendicular to the direction of propagation18. In level measurement technology. Radar-Füllstandsmesssysteme 5. If the direction of polarization coincides with this. Radar handbook 26 . 5. since there only such H vectors exist that run parallel to the conducting surface. strictly speaking.4 Polarization In the microwaves emitted by the antenna the vectors of electric field strength E and magnetic field strength H are constantly oriented (linear polarization). so that the resultant directional pattern can be represented in a two-dimensional graph19. 18 19 This applies to plane wave fronts and so. only to the far-field region of the antenna. In a first approximation it can be assumed that the characteristic is rotationally symmetrical around the main radiation direction. and also E vectors that are vertical to the conducting surface. 22. strong reflections are obtained from the wall and signals are cancelled due to wave interference. it can be seen that apart from the major lobe there are also side lobes which are particularly well developed for the dielectric rod antenna. the direction of the linearly polarized wave can be significant in the vicinity of metal surfaces (e. a tank wall). or the direction of polarization rotates in space and time (elliptical or circular polarization).

amplitude /[dB] -10 Horn antenna D = 6 x lambda -15 -20 -25 -30 Angle / [°] -90 -80 -70 -60 -50 -40 -30 -20 -10 0 0 10 20 30 40 50 60 70 80 90 -5 Rod antenna L = 12 x lambda rel. amplitude /[dB] -10 -15 -20 -25 -30 Angle / [°] Fig. 22: Measured directional patterns of a horn antenna and a rod antenna Radar handbook 27 . Radar-Füllstandsmesssysteme -90 -80 -70 -60 -50 -40 -30 -20 -10 0 0 10 20 30 40 50 60 70 80 90 -5 rel.5 3.

Normally.4) and thus to signal distortion.4. 10 bar/145 psi and higher) b) waveguides used as stillwells. the influence of the atmosphere on the speed of light need not be taken into account for the measuring accuracy of microwave level measuring systems.1 Influence of the carrier medium (atmosphere) Microwaves are propagated almost independently of the carrier medium. permittivity of gas at normal conditions normal temperature (in Kelvin). Exact compensation. the free-space speed of light in a vacuum or in air is generally taken as the basis. generally 105 Pa (= 1 bar = 14. The corrected distance a is then: a = ao · K The following subsections describe the correction factor K = c/co obtained from theoretical calculations.6 6. for example. however. generally 273. pressure and temperature: where: εr. Correction is relevant under the following conditions: a) high pressures in the tank (with air. Such a systematic error of the measured distance can be taken into account by using a correction factor K in signal evaluation.15 K (= 0°C/32°F) normal pressure.1 Propagation rate To calculate distance with the radar measuring system. since multi-modal propagation within the waveguide would lead to different transit times (see chapter 4.1. It is important that only the desired fundamental mode be generated in the stillwell. the greater is the risk that higher modes will be activated by interference. 6. but deviating propagation rates can occur in special application conditions. The greater the diameter of the stillwell. approx. This method is normally adequate. Wave propagation 6. requires a calibration measurement.N TN pN T p rel. It is very close to unity but is dependent on medium. The relative permittivity εr of the gas in the atmosphere above the liquid determines the propagation rate of the microwaves.5 psi) temperature (in Kelvin) pressure Thus the correction factor as the ratio of propagation rate c to the speed of light in a vacuum co is: 28 Radar handbook .

58 · 10-3 -3 Medium Air Carbon dioxide Hydrogen chloride Ammonia εr.55 · 10-3 0.60 · 10-3 7. nitrogen and argon behave similarly to air. Radar-Füllstandsmesssysteme For air at normal conditions.996 0.997 0.52 · 10-3 0. Oxygen.987 0 Normalized propagation rate in air temperature 150°C / 300°F 100°C / 210°F 150°C / 120°F 125°C / 75°F 110°C / 30°F -25°C / -15°F 5 pressure 10 145 15 20 290 25 30 435 35 40 bar 580 psi Fig.1 0. dependent on εr.990 0.993 0.20 · 10-3 Radar handbook 29 .994 0.6 3. the measuring error then exceeds 0.N . 23: Change in propagation rate at pressure and temperature Significant deviations arise only when pressure levels exceed 10 bar (145 psi).991 0.992 0.N . For deviating pressures and temperatures the following graph is obtained: K 1.998 0.989 0.988 0.3%.03%.995 0.26 · 10-3 0.000 0.1 0.59 · 10-3 1. with other gaseous media the quantitative effect can. A correction will of course only make sense provided the pressure in particular remains approximately constant during measurements.N be lower or even substantially higher: Medium Helium Hydrogen Oxygen Argon Nitrogen εr.999 0.07 · 10 0.00 · 10-3 4. the difference in the propagation rate compared to a vacuum is only 0.

1.g. 24: Change of propagation rate in the waveguide (K = c/co) diameter / wavelength D / 20 Heat loss in the conductor / dielectric material Radar handbook 30 .5). The liquid surface to be measured is located inside the stillwell.4).1. the propagation rate is: Given a spatially restricted dielectric (e. 6. or the coaxial cable is filled with a dielectric.6 3.4.4. Radar-Füllstandsmesssysteme 6. value c/c0 needs to be specially calculated.3) or given ohmic and/or dielectric losses20.4.3 Propagation rate in stillwells/waveguides Another physical factor affecting the propagation rate is when microwaves are propagated not in free space but inside a pipe acting in the same way as a waveguide (see chapter 4. planar line.2 Propagation rate on lines Electromagnetic waves are propagated at the speed of light along a loss-free line (sections 4.4. Correction factor for propagation rate in the waveguide Fig. section 4. When the line is completely surrounded by a dielectric εr.1 to 4.

for other modes the factor 1.4. The propagation rate in the cylindrical waveguide (inside diameter D) is22: The diagram in Fig. 25). 24 shows the correction factor as a function of the inside pipe diameter.5. incident wave transmitted wave reflected wave dielectric window Fig. 6. The minimum diameter is dependent on the wavelength. glass. We therefore obtain the ideal thickness d of the window with material property εr as: 21 22 Relevant is the so-called group velocity. 25: Reflection of waves from a dielectric window The series of events can be imagined thus: a first (negative) reflection r1 occurs at the left interface between air and dielectric and a second (positive) reflection r2 at the right interface.6 The thinner the pipe. It is reasonable to assume that the difference in the transit time of the waves must be a multiple of the wavelength inside the dielectric. To this end. 31 Radar handbook .7 will change according to the table in section 4. the formula applies to the fundamental mode of the H11 waves. "windows" of a dielectric material (plastics. ceramics) are used which for the microwaves should preferably be transparent and low reflecting.2 Transmission through dielectric windows The practical application of radar level gauging systems requires that the tank interior be separated from the tank exterior in order to provide separation of pressure. divided up into incident wave and reflected wave. the slower are the waves propagated21. which ideally cancel each other out (Fig. or frequency. temperature and product. not the phase velocity of the waves Group velocity.

the power density is largely constant23 (independent of a): 6.02 … 0. Radar handbook 32 . slight attenuation does occur due to the resistance of the pipe wall. the equation is simplified to: P’[dB] = P[dB] . for stainless steel waveguides about 0. An isotropic (having the same physical properties in all directions) radiator distributes its transmission power at distance a on a spherical surface of size 4 π a2. typical values for copper are about 0. Including antenna gain G. 23 However. α is the relevant attenuation factor (unit m–1 or ft –1): P’ = P · e – α · a If the powers are given in dB and attenuation in dB/m or dB/ft.2 … 1. the following is given for the power density p: There is no free-space attenuation in a stillwell.a[ft] · α [dB/ft] Attenuation must be considered for both ways.5 dB/m.6 3.4 Atmospheric signal attenuation Given transmission losses. power additionally decreases exponentially with distance a.a[m] · α [dB/m] P’[dB] = P[dB] . Radar-Füllstandsmesssysteme 6.3 Free-space attenuation By this is meant the power density decreasing in step with increasing distance from the radiator in the loss-free medium.2 dB/m.

According to Section 6. in practice. 6. The case is different with liquid ammonia (NH3). However. Fig. The absorption maximum at approx.7 the relations for antenna gain and propagation loss can now be included. 26: Attenuation of microwaves in air.3 the power density of the waves before reaching the reflector at distance a is given by: The distinction now needs to be made between 2 cases. see Section 7. 20 GHz is due to water vapour. which under pressure (approx.10 bar/145 psi at 10°C/50°F) usually forms a dense gas phase above the liquid. so that in the ideal case24 the transmission power is totally reflected.1. temperature and relative humidity. the total path a + a is travelled to the receiving antenna. in practical terms microwave attenuation in air is not significant for radar level measurement systems because at typical tank heights of up to 30 m (100 ft) it is at most only about 15 dB/km · 2 · 30 m = 1 dB. Therefore.5 g /m2 moisture content. and the power density is: 24 Reflection factor R = 1.6 The attenuation factor is dependent on frequency. the maxima at 60 GHz and 120 GHz to oxygen. 26 for air at 20°C and 7. Attenuation performance is dependent on pressure. depending on the size of reflecting area AR: (a) The reflecting area is so large that it intersects the beam cross-section completely. It is given in Fig. 33 Radar handbook .5 Modified radar equation Based on the radar equation given in Section 3. it dampens microwave signals in the X-band (10 GHz) so much that no reflection is measurable.

is: By multiplying the power density with the effective antenna area and the receive efficiency η2. so that the power density at the receiving antenna. depending on the size of reflecting area AR: (a) The reflecting area is so large that it intersects the beam cross-section completely. located at distance a from that reflector. Therefore. According to Section 6.5 Modified radar equation Based on the radar equation given in Section 3. Area σ (see Section 7. the effective radar cross-section of the reflecting target σ is relevant. the total path a + a is travelled to the receiving antenna.5) then acts as an isotropic radiator with power p1.7 the relations for antenna gain and propagation loss can now be included. see Section 7.3 the power density of the waves before reaching the reflector at distance a is given by: The distinction now needs to be made between 2 cases. Radar handbook 34 . so that in the ideal case24 the transmission power is totally reflected.6 6. and the power density is: (b) When AR is smaller than the beam cross-section. we obtain the received power: 24 Reflection factor R = 1.σ.1.

atmospheric attenuation and effective reflector area are independent of frequency. This applies only up to a point to the cross-section of the reflected beam σ. since e. antenna. 27 gives a comparison of various radar systems using different antennas and transmission frequencies. reflection factor.8 GHz system with 130 mm (5”) dia.8”) dia. and atmospheric attenuation α and a reflection factor R of the reflecting surface (see Section 7). for a plane face it is proportional to f 2 (see Section 7. Where a level measuring system is used in a large-area tank. or a 10 GHz system with 100 mm (4”) dia. 27: Transfer function of radar systems using different frequencies and antenna sizes (normalized to D = 100 mm (4”) at 10 GHz) The distinct increase in power due to higher frequencies and larger antennas is quite evident. Normalized transfer functionn 100 f=50GHz 10 f=24GHz f=10GHz f=5. there is the same proportional dependence25 on antenna diameter D and wavelength λ or transmission frequency f: Fig. it is better to use equation (b): decrease with the 4th power of a.0. equation (a) should be applied by approximation: the signal decreases to the square of a. however.1 (4") 0.8GHz 1 0. In both cases.6 Also taking into account the equation for antenna gain G1 (Section 5.1 0 20 60 80 100 120 140 160 180 (8") 200 antennea diameter [mm] Fig.2). 35 Radar handbook . A 50 GHz system with a 45 mm (1. Where extremely tall tanks or interference reflections from small internals are involved. antenna would have the same power as a 5.5). 25 Assuming that efficiencies. we obtain the following interrelationships: A significant point is that the received power decreases with increasing distance a.g.

In practice.m.6 Equivalent isotropic radiation power (EIRP) The equivalent isotropic radiation power (EIRP) is calculated In order to allow assessment of the effective radiated power in the main-radiation direction. 28: Arrangement to measure the EIRP When allowing for the loss in free space Dfree = (4πa/λ)2.s. an EIRP = -45 dBm at a distance of 3 m (10 ft) corresponds to a field strength of approx.6 3.s.m. 36 Radar handbook . the EIRP is then calculated by: In turn. whereby the distinction has to be made between the peak value Ep and the r. the electric field strength E. the received power PE is measured at a defined distance a by means of a reference antenna (gain G2): Fig. Radar-Füllstandsmesssysteme 6. can be calculated as follows. value Eeff of the field strength: For example. 460 µV/m peak and 325 µV/m r. dependent on distance a. This is equal to the product of transmitted power PS and antenna gain G1: EIRP = PS · G1.

1 to 4. resp.. 30. 30: Pattern of electric field lines in the plane across the line: a) single-wire. moreover. Distance Useful signal Fig. 37 Radar handbook . Signal level Interference signal An interference reflection is generally formed at the junction between measuring system and rod (see chapter 4. If in doubt. 26 As the skin effect is heavily dependent on the magnetic property µr.4.6 6.7 Propagation along electric lines (TDR method) In the TDR method. but this only becomes apparent in steel lines of more than 10 m (32 ft) in length26. c) coaxial cable.8). On the basis of the field patterns for single-wire. waves are propagated not through the tank atmosphere but along an electric line (see chapters 4. have a conductivity that is approximately 40 and 25 times. copper or aluminium should be used. 29 by way of example with a rod or wire rope. higher than steel. shown in Fig. two-wire and coaxial cables shown in Fig. it will be recognized that the possibility of interference from adjacent metal objects is the greatest with the single-wire system and non-existent with the coaxial system.1.4. b) two-wire. both of which. Line attenuation occurs with the finite conductivity of the metallic material.4. a completely non-magnetic (austenitic) steel is to be favoured.4).2). The propagation rate along the line is equal to the speed of light. independent of the type and dimensions of the line (see chapter 6. 29: TDR arrangement with a rod Fig.

g. reflection is almost 100%: R = 1. so that its influence is negligible. Electromagnetic waves are reflected by electromagnetic interaction: a) from conductive surfaces (metals. 31. error probability and detectivity in the case of non-ideal surfaces or interference reflectors. see Fig. repeatability.1 Reflection factor Significant operating parameters of a radar level measuring device are dependent upon the reflected useful signal of the microwaves. the relative permittivity is a complex number εr = εr’ + jεr” whose imaginary component εr” quantifies the loss (damping) in the dielectric. The power reflection factor R is defined here as being the ratio of reflected power density to the power density of the incident beam R = prefl /p. which describes the interaction with electric fields27): the strength of reflection is a function of εr28: At a relative permittivity εr = 3. b) from dielectric liquids (described by the relative permittivity εr . accuracy. the absolute value of εr can by approximation then be taken for calculations. Generally speaking.5 about 10% of the signal power (–10 dB). Reflection factor as a function of the relative permittivity dB R Fig. In the frequency range > 1 GHz under consideration. alcohols. 31: Reflection factors of dielectric media 27 28 There is an additional dependence on the relative permeability µr which describes the magnetic behaviour of the medium. Exceptions are for instance water. the imaginary component is close to zero for most liquids. however µr is very close to unity. the relative permittivity plays a central role in the evaluation of reflectivity and applicability of the microwave level measuring system. Reflection 7. nitrobenzene. e. For almost all materials. and highly conductive liquids such as acids and saline solutions of sufficient concentration). and at εr = 1. In these cases. measurability.5 only 1% of the signal power (–20 dB) is reflected.7 7. Radar handbook 38 . Hence.

32: Reflection factors at a liquid-liquid interface Radar handbook 39 . Reflection factors with 2 liquid layers Fig.2. and also the reflection factor R2 at the interface for various parameters εr. the reflection factor is: If the power received from that point in relation to the transmitted power is to be calculated.g. It will be seen that the interface reflection becomes stronger the greater is the difference in relative permittivity.1 and εr. Therefore. problems often arise because some products with a high εr (e.2.1 >> εr. Even where εr. it must be borne in mind that the waves still have to travel twice through the transition zone between the atmosphere and the upper layer with in each case a transmission factor of (1-R).1. however.7 3. In practice.2 the interface can in theory be readily detected.2 Reflection at interfaces When microwaves strike an interface between 2 media with relative permittivities of εr. Radar-Füllstandsmesssysteme 7. water) absorb microwaves. The following diagram shows the reflection factor R1 at the upper liquid layer as a function of εr. the effective reflection factor R2 at the interface is: provided that the waves are propagated loss-free through the upper layer.

have high εr values. but these probably decrease for long-chain compounds. aldehydes and ketones.85 · 10–12 As/Vm.3 Dielectric permittivity The (relative) dielectric permittivity29 is a dimensionless quantity describing the physical behaviour of a material in an electric field. This also applies to the sulphur/oxygen componds (sulphuric acid. Radar-Füllstandsmesssysteme 7. For vacuum conditions. for gases only negligibly greater than 1. εr = 1. even if this term is still often to be found in the literature for εr. oxygen. have an εr that is only slightly greater than unity. alkenes) or mixes thereof (gasoline.1 Chemico-physical assessment The value of the dielectric permittivity is dependent on the electric dipole moment of the compound. With diatomic elemental gases (e. which in turn depends on the molecular geometry and on the electron distribution. helium) which are monatomic. oils) have an εr = 2.7 3. alcohols. fluorine). ● Noble gases (e. sulphuric dioxide). for liquids usually much higher (normally ≥ 2). hydrogen and oxygen) and also an asymmetrical structure (e. The following describes the dependence of the permittivity on various influencing factors. ● ● ● ● ● 29 The dimensionless quantity εr should be termed relative permittivity and not dielectric constant.g. 7.g. On the other hand.5. water and ammonia) have a high εr . the following applying to the free space: ε = εo= 8. nitrogen. the rotational symmetry is only slightly disturbed. for water very high.g. Because of their slight asymmetries. εr being about 1. due to affine OH and O groups.g. The dielectric constant ε is the dimensional physical material constant with ε = εr·εo. with εr = 80. Carbonic acids are strongly polarized only in short chains. Radar handbook 40 .3. Inorganic compounds that feature atoms with different electron affinities (e. and only then have a high εr. the basic hydrocarbons (alkanes.

e. with the 3 positional isomers having εr values of between 2. Radar-Füllstandsmesssysteme Cyclic compounds because of their high symmetry of faces have lower εr values. and depends essentially on the symmetry in the molecule. A table with εr values of different products can be found in Annex A.e. 33: Dielectric constant of water as a function of frequency [Meinke] In this transitional range the imaginary component of εr is relatively high.e. which has at different temperatures the complex dielectric constant εr’+ jεr’’. For surface reflection. Due to dielectric relaxation (i. there is a transitional frequency range30 in which εr drops. shown in Fig. 3332. these frequencies lie between a few 100 kHz and a few 100 GHz. wave attenuation would also occur in the medium. i. Above this range.3. this range is however located just inside the 5-100 GHz microwave range. this has no negative influence. the electrically polarized molecules cannot align themselves fast enough in the highfrequency field). A further drop in the εr due to ion or electron resonance occurs only far above the microwave range. Dichlorobenzene is a case in point. 7. εr remains constant 31. With a few media. The dielectric characteristic of nitrogen or halogen derivatives varies considerably. For most liquids. 30 31 Radar handbook 41 .5 and 10.2 Frequency dependence The permittivity εr decreases in step with increasing frequency f.g. Fig. A good example is water. however. 32 The imaginary part jεr” describes the dielectric losses in the dielectric.7 ● ● 3. in the infrared or visible light range.

D) Media with low εr < 3. Radar-Füllstandsmesssysteme Other media. εr drops at higher temperatures (typical values are around approx. Basically. 7. – 0. it must be assumed that the εr is at least equal to the lowest εr of the constituents. NaOH. which already have a low εr at low frequencies. aqueous solutions of acids. media can be divided roughly into four categories: A) Media with high εr > 20 (tabular data for low f ). 3-6) is constant at f > 5 GHz. despite containing a large proportion of water. bases and salts.7 3. NaCl. the lower the transition frequency. C) Media with relatively low εr = 3-6. with reference to their permittivity and frequency dependence. sulphuric acid or acetic acid). and the transition frequency is usually lower. Also. dropping only slightly in the microwave range. have an εr that is quite distinct from that of water (εr = approx. changing below the micro-wave range so that εr (approx.3. constant up to and into the microwave range. B) Media with high εr > 20 (tabular data for low f ). maintain this value up to microwave frequencies and beyond. 42 Radar handbook .4 Liquid mixes Given a mixture of two or more liquids.3. 20 … 30 for solutions of ammonia.1%/K) On transition from the liquid to the solid state of aggregation the εr generally drops abruptly (ice for example has an εr of only 3. i. Examples: Water: Organic liquids: εr drops at temperatures below 25°C Mostly negative temperature coefficient.3 Temperature and viscosity dependence A definitive temperature dependence cannot be specified as behaviour differs from substance to substance. changing in the microwave range to εr > 9 at 10 GHz. 7. An approximation equation for a mix of two liquids with a1 and a2 parts by volume reads: In(εm) = a1 · In(ε1) + a2 · In(ε2) Organic liquids with a low εr and an added low water content have practically the same εr as the pure organic substance. there are initial indications that the transition frequency is dependent upon the viscosity of the medium: the higher the viscosity. On the other hand.2).e.

4 Scattering from particulate materials Depending on the relationship between particle size (diameter D) and wavelength λ.01·L[%] ).7 3. D << λ: The very fine-sized surface acts in the same way as a liquid. a material with εr = 2 and 50% air content has an effective εr . λ/4 to 3λ) heavy scattering must be expected.eff of 1. 34: Loss of reflection from particulate materials 34 Also to be considered is the reflection factor due to the εr of the particulate material (incl. Radar-Füllstandsmesssysteme 7. The rule of thumb: εr.5 Bulk (particulate) materials The effective relative permittivity of particulate materials (with air in the interstices) can be much lower than that of the homogeneous solid body.5. Also to be borne in mind is the fact that with bulk materials having a particle size range in the order of the wavelength. by approximation the following applies to the diffuse reflection [Ries] (see Fig. For instance. L = volumetric air content in %.3. D ≈ λ: If the wavelength is in the same order of magnitude as the particle size of the bulk material (approx. 7.eff = 1 + (εr –1) · (1 – 0. Practically no reflection can be measured. the microwaves will disperse (see next chapter) and so greatly decrease reflectivity. 34): · Reflection loss from particulate materials Fig. the waves are reflected or scattered when meeting the surface of particulate materials34: D >> λ: The surfaces of the particles act like miniature reflectors that reflect the waves according to the cross-section of their reflected beam. 43 Radar handbook . air).

g. Radar-Füllstandsmesssysteme 7. e. the radar cross-section σ must be considered as the effective reflection area. This particularity can be significant if the antenna axis (assuming horizontal surface of the medium. In general. since the relevant radar equation (see Section 6. as a rough estimate the maximum allowable angle of deviation between the direction of propagation (antenna axis) and the normal line of the surface is equal to half the lobe angle (Fig. 35 centre. it can be calculated as [Baur]. in the case of particulate materials or due to a running agitator). 35 left).5 Reflected radar cross-section from limited targets If reflection does not take place over the entire beam cross-section. The optimum arrangement is orthogonal alignment of the antenna axis and reflection area. at an angle opposed to the normal. a liquid) is not vertically aligned (Fig.5) uses a sphere as reference for the secondary radiator. 7. right: still acceptable angle of deviation 44 Radar handbook . centre: inclined reflection area. [Philippow. pipe) σ = 4π·A2 / λ2 σ = 4π·b4/3λ2 σ = π·r 2 σ = 2π·/2·r /λ A = area b = side of triangle r = radius r = radius.7 3. exposed to radiant radiation (e. Fig.4]: Large flat plate (any shape) Triple reflector Sphere Cylinder. whose dimensions are large in relation to the wavelength. or the reflection area is oblique (Fig. For some geometric bodies. l = length The radar cross-section can be substantially larger than the actual reflection area. that are large as compared with the wavelength λ.g. 35: Influence of the angle of reflection in the case of non-orthogonal arrangements: left: inclined antenna position.6 Angle of reflection The waves are reflected from a surface. 35 right).

is given by: ∆a = c · τ / 2 By virtue of the analogy between pulse duration and bandwidth B (see Section 3. 36: Reflected signal at different pulse lengths t Object discrimination.1) the following general “uncertainty relation” can be set up between B and the object discrimination ∆a. transmitted pulse 1st object 2nd object t Fig.5. this requirement is normally not important (amax= 30 m → fi ≤ 5 MHz). the signals cannot be unambiguously assigned. Radar handbook 45 . Hence it is necessary to satisfy the following relation (f i = pulse repetition rate): fi ≤ c / 2 · amax For the relatively short distances involved in level measurements. 36.8 8. however. i.2 Unambiguity If another pulse is emitted before the reflected signal from the previous pulse has been received. the two reflections will not be separable (Fig. However. the minimum difference in distance allowing two objects to be distinguished. bottom). the system will also receive a corresponding number of pulses (Fig. 36. top). if the pulse length τ is too long.e.1 Object discrimination If two or more objects reflect the radar signal. Evaluation methods 8. which also applies to FMCW systems: ∆a = c / 2 · B 8.

Even if digital evaluation by means of discrete frequency transformation (section 8. the signal frequency can nearly always be accurately determined36.2) is applied. ∆t is the repeatability that is determined by the jitter 35 error of the sampling. the uncertainty of time ∆t influences the result: ∆a = c · ∆t / 2 Neither is there any change when an extended time scale method is used (cf.8.2]: ∆a/a ≤ 8 · ∆ F/F Independent of the radar method used. In regard to level measurement. 35 36 Irregular fluctuations of the sampling points of time. no significant value can be determined for the level. Radar handbook 46 . measurements must be blanked out for such times. ● Interference reflections Various internals (pipes. vapour.1). The error is [Stolle. the signal frequency can be calculated exactly with interpolation techniques. In this case.g.8.4 Interference Various forms of interference can falsify the received radar signal in relation to the ideal reflection pattern. an appropriate (error) message must be available. 37): ● Atmospheric effects: Heavy damping or scattering from particles in the atmosphere (dust. 8. etc. They need to be given consideration and if necessary included in the signal evaluation in order to avoid misinterpretation. they may be included in the signal evaluation (see Section 8. level below agitator).6). Radar-Füllstandsmesssysteme 8. → If reproducible.7. foam. other sensors. etc.g.3 Measurement uncertainty The measuring accuracy of a radar distance measurement is determined by the uncertainty of time measurement. filling nozzles. interference reflections influence the measuring accuracy because of interaction when they occur in the vicinity of the useful signal (see also section 8. but the relative measurement uncertainty is influenced by the linearity of the frequency sweep ∆F/F. significant interference factors are (see Fig. condensation or deposits on the antenna) can also produce reflection signals. In FMCW radar. section 8.) or mediuminduced interference (e.8 3.) → If the surface of the medium is no longer detectable. agitator blades. if the surface of the medium is at times obscured (e.1 “empty-tank spectrum”). For the pulse radar. However.

8 3. for example. Other microwave transmitters: Several radar systems that are installed in one tank can mutually influence one other. the reflection signal is thus broadened in time and the measuring accuracy reduced → The antenna should be moved further away from the wall. this probability is normally very low because the systems would have to operate in synchronism down to fractions of µs in order to generate an additional differential frequency portion within the processing bandwidth of a few kHz. A better solution is to change the mounting position so as to eliminate multiple reflections altogether. and is again reflected from the medium before being received by the antenna → Since multiple reflections occur at periodic intervals. 37: Possible sources of interference in radar level measurement ● Multiple reflections: These occur. Multipath propagation: If. a signal is deflected from the tank wall. then strikes the tank cover or some other “good” reflector. In pulse radar with a high pulse repetition rate. ● ● Radar handbook 47 . for example. an interference can however easily occur when the signals from several transmitters are interpreted as being the total reflection pattern. its propagation path is lengthened. however. Radar-Füllstandsmesssysteme 2nd transmitter internals multiple reflection multipath propagation agitator Fig. they can be detected and taken into account in the signal processing. With FMCW radar. when the signal is reflected from the surface of the medium.

03 = –15 dB.5 Sample calculation for received powers An example is given below for calculating the received power of a radar system.7 kHz. 8. a = 20 m (60 ft).5. In accordance with the formula given in Section 7.1. signal frequency is f = 6. i. the receiver noise power thus amounts to: P‘noise = 4 ·10–16 W = –124 dBm. 8.5. Let the transmission power be PS = 1 mW = 0 dBm.5. The receiver bandwidth is used for the effective bandwidth B. Let the tank diameter be such that a large reflecting surface and case (a) from Section 6. so that the wavelength is λ = 3 cm (1.2: P‘noise = FkTB.5. 48 Radar handbook . let a tank contain a poorly reflecting medium (gasoline.5. the signal-to-interference ratio amounts to only 2 dB.3 Signal-to-interference ratio of an interference reflector By way of example it is assumed that a metal plate (R = 1) with an area A = 10 x 10 cm2 (4 x 4 in2) is located at distance a = 10 m (30 ft) below the antenna.4 m2 The received power is calculated according to Section 6.2”). case (b): In this example. Thus the antenna gain in accordance with Section 5. Radar-Füllstandsmesssysteme 8. The signal-to-noise ratio is: SNR = 74 dB. Atmospheric damping may be disregarded (α ≈ 0).1 Useful signal For ’worst case’ estimation.5. the received power is given as: 8. Normally noise is not a factor in radar level measuring equipment.3. and the signal-to-noise ratio and signal-to-interference ratio for an interference reflector. At T = 300 K. the radar cross-section of the reflector is: σ = 4π · A2 / λ2 = 1.2 is: Let the tank height be 20 m (60 ft).7.e.8 3. B = 10 kHz and a noise figure F = 10 dB are assumed. Therefore. According to Section 6.5 can be assumed. Let a horn antenna with D = 200 mm (8”) have an efficiency η = 0. As given in Section 7.2 Signal-to-noise ratio The noise power of the receiver as given in Section 4. Let a 10-GHz radar be used. In an FMCW system with 1 GHz sweep and 20 ms sweep time. εr = 2) at a low level. the reflection factor of the medium is thus: R = 0.

8
3. Radar-Füllstandsmesssysteme
8.6 Signal evaluation in the pulse-radar system Both in pulse radar (see 3.5) and in the TDR method (see 3.3), which normally use pulseshaped signals, the signal propagation time must be evaluated directly. This lies in the nanosecond range, and the necessary resolution in the picosecond range. A so-called extended time scale method is used to be able to measure such short times. Fig. 38 shows the elementary circuit arrangement consisting of two oscillators which oscillate with a slight frequency shift Df, stabilized by a PLL. Accordingly, the signal received via a directional coupler is sampled at every following pulse with a time delay of:

The time scale factor is thus f0 /∆f.

Sampling Pulse shaper

Directional coupler

Sensor Transmitted pulses (fo) Received pulses Sampled pulses (fo-∆f) Extended time signal (ZF) Fig. 38: Block diagram of an extended time scale circuit and signal diagram

The extended time signal can be evaluated by digitizing and subsequent valuation of the reflection characteristics (determination of maxima above an envelope).

Radar handbook

49

8
8.7 Signal evaluation in FMCW radar In FMCW radar, the information on the measured distance is contained in the frequency of the low-frequency signal obtained from the mixer. Accordingly, various methods can be applied to determine the frequency of analog signals, and these are outlined in the following subsections. 8.7.1 Counting method The simplest evaluation method consists in measuring the time interval between two zero crossings of the oscillations (Fig. 39: e.g. t1-t0 , t2-t1…), Results are more accurate when a greater number of cycles N are counted: ∆t = (tn-t0) /N. The problem with this method lies in the fact that the signal-to-interference ratio must be reasonably good in order to avoid making errors when determining the zero crossings. Given large interference components, the method is no longer suitable.

time t

Fig. 39: Counting method to determine frequency

8.7.2 Fourier transform The method commonly used for signal evaluation provides for a Fourier transform with the aid of digital signal processing. The signal is first digitized by being sampled at constant intervals. This is followed by discrete Fourier transformation (FFT 37) into the frequency range (Fig. 40). Generally several adjacent spectral lines occur 38. The spectal lines formed by discrete Fourier transform of a signal (cf. Fig. 30) sampled at intervals T/N have a frequency spacing of: ∆f = 1/T. In accordance with the relation for the mixed frequency:

we obtain for the local spacing of two adjacent spectral lines:

37 38

Fast Fourier Transform Due to the finite scanning period and convolution with the time window, a sinc (= sin(x)/x) function is obtained as the envelope of the spectral lines. And in general, the signal frequency is located between the discrete lines. Radar handbook

50

8
3. Radar-Füllstandsmesssysteme
T
transmitted frequency

F
time N sampling points

(voltage values)

mathematical transformation

(line height)

FFT
time T 0 N time signal (behind mixer) 1 0 N frequency spectrum

Fig. 40: Digital signalprocessing in FMCW radar

Example: at a frequency sweep F = 1 GHz, a line spacing in the FFT spectrum can be calculated of 15 cm (6”) measuring distance. Using a method to locate the “weighted average” in the spectrum by interpolating between the discrete lines, however, the measurement resolution can be significantly increased provided only one reflector is located within a distance of less than ∆a. It is important to know that line spacing ∆a is only dependent on sweep F. It is not possible to improve object discrimination by changing the sampling frequency or the number of sampling points39. The above relationship is identical with the relation for the pulse radar system (see Section 8.1), if bandwidth B and sweep F are equated. The uncertainty relation formulated in 8.1 is universally valid. A great advantage of FFT evaluation is that useful signals and interference signals, provided their frequencies are far enough apart, are distinctly separated – even when the interference amplitude is greater than the amplitude of the useful signal. However, through superposition of close-by interference frequencies, it is possible for measurement deviations to occur which become apparent by a locally periodic error function40. 8.7.3 Phasegradient method As described in Annex B, the individual sampling points in the FMCW method contain phase information on the reflection signal at different frequencies. The phase gradient function can be calculated with the aid of the so-called Hilbert transform from the time signal, and from that the distance value. The calculation is more time consuming than a simple Fourier transform, but the result is more accurate when there is only small signal interference41. However, the phase gradient method is not suitable when interference signals are greater than the useful signal.

39

Even though there are methods of reducing the line spacing in the spectrum by adding further synthetic sampling values, they have no effect on the object discrimination. 40 Caused by superposition with the “sidebands” of the interference spectrum, which in accordance with a sinc function are located around the main frequency. 41 For example, by near interference reflections or frequency-dependent amplitude modulation of the RF Radar handbook 51

41: A high-pass filter and its filtering characteristics In all. from which in the fourth step the deviation of the estimated from the real value is calculated. but that a finitely steep filter slope is obtained. Fig. the higher signal frequencies will be reflected from half the sampling frequency and will produce spurious frequency contents after conversion and Fourier transform (Fig. such electronic filters are easy to set up and reproduce. 42: Visualization of frequency reflection when Shannon´s condition is ignored The antialiasing filter must be a very steep-sloped low-pass filter (at least 4th order) and is usually dimensioned for a cut-off frequency of 0. Fig. see Fig. for which e. it is possible by appropriately filtering the frequency to considerably increase the effective dynamic performance of measurement and thus improve signal quality. but the computational effort is substantially higher than when using the FFT.. the frequency spectrum f has to be limited to half the sampling frequency fA : f < fA / 2.8 . a signal is synthetized from the frequency and. If this is ignored. 0.7. 42): f’ = fA . The corrected frequency can then be used as the starting value for the next measurement. the following filters can be used: Antialiasing filter Due to time-discrete analog/digital conversion of the signal and because of Shannon’s sampling theorem. 52 Radar handbook . the FFT analysis can be used. Owing to the low signal frequencies involved. the circuit arrangement of a second-order high-pass filter with an operational amplifier.8 8. in the second step.. If the frequency value has not changed by too much between two measurements. the change in frequency – and thus the change in level – can be very accurately tracked. 8. 41. It is carried out in four steps: first the frequency is estimated. Hence the term “tracking”. Appropriately. by way of example. This comparison supplies an error value.4 Tracking The objective of this method is to determine the frequency of a digitized signal.5 Signal filtering Since in FMCW radar the information on the target distance is to be found in the frequency of the down-converted received signal (see Fig.7.g. in the third step. tracking is also carried out by means of digital signal processing.f Sampling frequency Frequency f Fig. compared with the measuring signal. A point worth noting for all practical filters is that the signals cannot be completely suppressed in the rejection band. 41 shows.95 · fA / 2. 8).

not only is the RF signal performance restricted but the dynamic performance of the measuring system is also limited. undesirably strong signals will in some applications occur in the case of long distances. which for instance stem from the mechanical tank separation system (see chapters 4. Adaptive filtering The recommended variable high-pass and low-pass filters can be most effectively used when their cut-off frequencies are set as a function of the measuring distance. by means of selectable resistors in Fig. a greater number of interference reflections occur in the case of short distances. the amplitude is inversely proportional to the frequency of the signal.e.4. and therefore low frequencies. such as when multiple reflections occur (see 8. signal amplitude diminishes with the square of the distance. This characteristic could simply be compensated by a double differentiator circuit which amplifies the higher frequencies accordingly. and assuming the same reflection conditions on the surface of the liquid. Signal strength (log. 53 Radar handbook . In practice.4).4. The variable clock frequency has to be substantially higher than the signal frequencies. reflection is located farther away the high-pass frequency can be amplified in order to reject short-distance interference reflections even better. 43) or even continuous variation by switched-capacitor filters (SCF)42. Thus. or wanted. so limiting the means of detecting very weak signals or very different amplitudes within a signal mix. filters can thus be implemented with a controllable cut-off frequency. high frequencies. Ideal is a circuit which allows changeover of the filter cut-off frequency (e. Since in FMCW radar the frequency is proportional to the distance.2) and the antenna. Such interferences can largely be eliminated by high-pass filtering. for example when the useful. a second-order high-pass filter is used with an appropriately high cut-off frequency (Fig. an equivalent resistance whose value is proportional to the clock frequency is synthesized by the periodic charging and discharging of a capacitor. i. Built into the negative-feedback branch of an operational amplifier.8 Spatial filter As explained in chapter 6. 43: Characteristics of the spatial filter High-pass filter In practice. which if necessary can also be of changeover design. Where level measurement is carried out by means of pulse radar. The filters are optimally adapted to the current measurement situation. None of these filters can be used in the pulsed radar system. 43). These interferences can be reduced by low-pass filters. 42 With the SCF.) Double differentiator (~ f2) Spatial characteristic (~1/a2 ~ 1/f2) High-pass filter Distance a Frequency f Fig.g.8 and 6. Low-pass filter Similarly.

45). however. Radar-Füllstandsmesssysteme 8. For current measurements. can be minimized by using the complex spectrum information (including phase) (Fig.empty tank spectrum 0 f = corrected spectrum Fig.1 Empty-tank spectrum Reproducible interferences caused by reflection points in the transmitter (transmission lines. absolute values may be used for the empty-tank spectrum. The method can also be applied by analogy to pulse radar. 54 Radar handbook . etc. tank fitting. Measuring errors occurring in the vicinity of interference reflectors. this “empty-tank spectrum” can then be subtracted from the established reflection spectrum.8. antenna) or in the tank (by fixed internals.8 Special methods 8. Fig. 44).8 3. 45: Measuring errors caused by an interfering reflector: evaluation with the absolute-value emptytank spectrum (left) and complex emptytank spectrum (right) Error [mm] (Absolute value empty-tank spectrum) Error [mm] (complex empty-tank spectrum) 500 Distance from disturbing reflector [mm] Distance from disturbing reflector [mm] The term “empty-tank spectrum” originates from applying the described method to the signal spectrum of an FMCW system. tank bottom. 44: Subtraction of empty-tank spectrum 0 f In its simplest form. interferences are accordingly blanked out (Fig.) can be suppressed if the reflection signal is measured and stored when the tank is empty. in which case the “empty-tank time signal” needs to be subtracted. measured spectrum 0 f .

is evaluated and the true level established by way of the known reduced propagation rate of the microwaves in the medium: ● Whereas waves in the tank atmosphere of height a are propagated at the speed of light c. so the tank bottom is practically “visible”. permittivity = εr. Here the reflection from the tank bottom.8. Given low attenuation in the medium. Radar-Füllstandsmesssysteme 8. where they are reflected and pass through the medium and the atmosphere before reaching the receiving antenna. the waves are propagated down to the tank bottom. waves in the medium (rel. 46: Mechanism of tank bottom reflection and tracing Radar handbook 55 . and the apparent tank height hv greater than the true height h.8 3. shifted in the spectrum (FMCW) or in the time signal (pulse). only a small portion of the power is reflected from the surface. since the propagation rate in the medium is slower than in the atmosphere. while the majority penetrates into the liquid or particulate material. the tank bottom appears to be shifted downwards. ● Hence. signal signal a h hv r1 r1 r2 tank bottom r2 signal in empty tank signal in filled tank Fig.2 Tank bottom tracing In media with a low relative permittivity. the reflection r2 from the tank bottom appears on the time axis or in the spectrum to be shifted downwards. Tank bottom tracing is a special evaluation method to allow measurements to be carried out in such application conditions. However. height L) are propagated at a slower velocity v. The medium is “transparent”.

these measurement problems can be solved using a non-contacting radar system. Radar-Füllstandsmesssysteme ● The transit time in the medium is t1 = L / v. 47: Principle of operation of a wire-conducting TDR level gauge (b) t2 2xt 1 2xt 2 2xt 3 time 56 Radar handbook . t0 RFelectronics conducted wave 0 < t1 < t2 < t3 2-wire line (a) 1 2 Signal at the sensor: Fig.3). d and from that filling height L can accordingly be determined exactly: The method can even be applied when signal from the surface of the medium is no longer measurable.3 Interface detection The requirements for interface detection are described here by way of example in connection with a TDR system (time domain reflectometry.a) to true filling height (h . h and hv are known. but the advantage of guided waves is that signal weakening is not so great and interference from the tank geometry is largely avoided.8.8 3. The ratio of apparent “layer thickness” (hv .a) therefore corresponds to the ratio of the wave propagation rates: ● Where εr . see section 3. whereas for the same distance in an empty tank it would be t0 = L / c. In principle. 8.

which can serve as reference point. layer of emulsion). The reflected power in each case depends on the difference of the relative permittivities (see section 7. The polarity of the signal is reversed at each reflection from lower to higher relative permittivity.8 3. For a reliable measuring system. part of the wave is reflected back to the sensor (time t1). Inadequate object discrimination (see Section 8. Radar-Füllstandsmesssysteme Fig.2). Risk of product deposits on the lines. causing additional reflections. At each position at which the surrounding relative permittivity ε changes.g. water).g. the following situations need to be considered for signal processing: ● Changing εr values due to changes in temperature or composition of product. No abrupt transition from one product to another (e.1) when thickness of the interface is too small. 2t2 and 2t3) indicate the positions of the interfaces and the end of the line. 47 shows that an electrical pulse is generated (time t0) and guided via a 2-wire line as an electromagnetic wave. ● ● ● Pronounced attenuation or absorption of the microwaves in the liquid (e. The wave is propagated along the entire line and is reflected a second time (t2) at the interface between the two liquids and a third time (t3) at the end of the line. ● Radar handbook 57 . The signal delay times (2t1.

elemental 1. tabular data [Weast].long-chain 2.anhydrides 20 c c c c c 2-9 (tr) ? ? 5 ? ? ? ? c c ? ? methane. fluorine .23 c hydrogen 1. propion-aldehyde acetone = propanone.05 c helium 1.8 Alkenes 2 Alcohols 30 14-20 . pentene etc.A Annex A Table of dielectric permittivity The following overview is based on data from the literature. ethane.long-chain (>C16) 2 .e.) . propylene.5 high f 44 εr ? 60 (tr) ? ? ? ? c ? Examples hydrogen cyanide HCN water H2O hydrazine N2H4 sulphuric acid H2SO4 ammonia NH3 hydrogen sulphide H2S Ge-.compounds 1. i.1]. vaseline mineral. butane pentane.5-2.1-2. tripene ethylene. [Hippel].liquids 2 .liquid gases 3 ..5 c argon.4-2. synth. Pb-. decreasing further at higher frequencies Radar handbook 58 . air 1.polyhydric 40 Aldehydes 13-22 Ketones 20 Acids & derivatives 58 6 3 . benzyl alcohol glycol. Si-. [VDI. Sn-. glycerol form-.1 c chlorine.10 GHz) ? = value not known c = εr constant into the microwave range tr = at 10 GHz still in the transition range. butanone formic acid acetic acid butyric acid fatty acids (various) acetic anhydride 43 44 quasistatic.5 . pentanol butanol.longer-chain 13-19 . decane. acet-. octane. gasoline paraffins. silicon oil. Ti-Cl4 Tetrachlorides Sulphur (liquid) Inorganic liquid gases (pressurized or at low temp.6 c carbon dioxide CO2 14 ? sulphur dioxide SO2 ORGANIC COMPOUNDS Alkanes . propane. application experience and laboratory measurements. methanol. nitrogen.oils 2. generally up to a few kHz in the microwave range (at approx. (εr values are rounded guide values for the media examples) Group of liquids INORGANIC low f 43 εr 115 80 52 22 17 9 2. oxygen.ester 3-16 .9 3. ethanol propanol.

1 .asiinstr.m-. PA. i. phenols trichloroethylene carbon tetrachloride. trimethyl-. for example.e.5-5. PVDF glass Al2O3 ceramics PVC powder alumina bauxite 3-8 2. Radar-Füllstandsmesssysteme (continued) Group of liquids Ether Cyclic compounds low. PTFE .5-8. generally up to a few kHz in the microwave range (at approx.5-3. chloroform chlorinated diphenyl = clophen chloroacetic acid acetyl chloride methylamine isopropyl-. decreasing further at higher frequencies 59 Radar handbook .5 2. diethyl-. PC.com/dc1. -hexene C6H10 toluene.html 45 46 quasistatic. cellulose nitrate.4 2 2. melamineformaldehyde.A 3.3 3. cellulose acetate.5-10 5-6 5. PP PS.10 GHz) ? = value not known c = εr constant into the microwave range tr = at 10 GHz still in the transition range. -phenol. tetrachloroethylene dichlorbenzene (o-.8-6. dioxan benzene C6H6 cyclohexane C6H12.5 A comprehensive table with many other products can.9-2. xylene etc.3 2 2.amides 60 Plastics (solid form) 1. f 45 εr high f 46 εr ? c c c ? c c ? ? c ? ? ? ? 30 (tr) ? ? ? c c 3 4-5 c c c c c Examples diethyl ether.7 4.5 3-3.2 2. nylon phenolic aldehyde.5-5 5-9 Solids Particulate materials 3.4 10 Halogen derivatives 3.4 1. amylamine nitrobenzene nitroethane aniline acetamide PE.5 35 28 5-7 .acid halides 33 16 Nitrogen derivatives 10 3. benzyl-. ABS PVC.p-) chlorobenzene. be found on the Internet under: http://www.

phase evaluation is made between the transmitted and the received signal. delayed with τ = 2a/c : ϕ ϕE ϕS f τ c a λ phase phase of the received signal phase of the transmitted signal frequency delay time of the wave propagation rate of the microwaves = speed of light distance of the reflector = c / f = wavelength (A1) A sine or cosine signal s can be described in general by: A ϕ f t ϕ0 amplitude phase frequency time start phase (A2) With a linear frequency-modulated signal the modulation constant m is included in the cosine function: (A3) The instantaneous frequency of signals with a time-dependent frequency is generally calculated by the differential of the phase: (A4) 60 Radar handbook .B B System-theoretical comparison between interferometer and FMCW method With the interferometer method.

Eq. (A3) with linear frequency modulation. Signal-theoretical description of the FMCW method: The phase difference of both signals (transmitted and received) is actually determined by the process of mixing two high-frequency voltages and subsequent digital sampling of the mixer output signal. Then the high frequency part (f1 + f2) is eliminated (e. by a low-pass filter): Behind the low-pass filter: The phase is: (A7) (A8) For the further considerations the signal amplitudes A1. Radar-Füllstandsmesssysteme For the signal in Eq.g.B 3. The FMCW system therefore acts in exactly the same way as an interferometer which determines the phase with the many different frequencies that are present at the respective sampling times. (A11)).g. (A3) can also be expressed as: (A5) (A6) For evaluating the phase of two signals s1 and s2 ideally a multiplication of the signals is performed (e. Except for a negligible error term (see Eq. the instantaneous (time-dependent) frequency is: F frequency sweep T sweep time Hence. A2 and the start phases ϕ1. ϕ2 are disregarded because they are constants and do not influence the results of the phase evaluation. Radar handbook 61 . by a mixer). it is immaterial in signal theoretical terms whether the frequency ramp rises continuously or in N equidistant steps [Stolle].

62 Radar handbook . 20 m) to give: At a half-wavelength λ/2 = 15 mm (f = 10 GHz). a calculation is made by way of a practical example for an FMCW radar method: F = 1 GHz.B 3. (A1). (A5). see Eq. the phase difference is: (A11) where f is the instantaneous (time-dependent) transmission frequency. Radar-Füllstandsmesssysteme The phase of the transmitted signal (see Eq.(A6)): (A9) The phase of the received signal i s: (A10) Therefore. Except for the error term F τ 2/2T the result is identical with the interferometer method. and is thus negligible. the measuring error then amounts to: ∆a = 4 ·10–4 · 15 mm = 6 µm. τ = 130 ns (equal to a measuring distance of approx. To assess this error. see Eq. T = 20 ms.

Schiek. Gundlach. Heidelberg. Boston. Beuth Verlag. The Record of the IEEE 1995 International Radar Conference.1] [Brumbi. Mikrowellentechnik.4] [Ries] [Salema] Radar handbook 63 . Erich. Eugen (Hrsg. Taschenbuch der Hochfrequenztechnik. G. Radarverfahren zur Füllstandsmessung in Großraumbehältern (Studie). Karl-Heinz (Hrsg. Salema. Brumbi. Arthur (Editor). 1982 Bonfig. London. Teubner. DIN-VDE 0848 Teil 2: Sicherheit in elektromagnetischen Feldern – Schutz von Personen im Frequenzbereich von 30 kHz bis 300 GHz. 1990 Brumbi.C C. Dielectric Materials and Applications. 1992 Musch. 1979 Ries. Technical characteristics and test methods for radio equipment to be used in the 1 GHz to 40 GHz frequency range.1] [Pehl.). Bibliography [Baur] [BG] Baur. Köln. Einführung in die Radartechnik. 1985 Berufsgenossenschaft der Feinmechanik und Elektrotechnik. Berlin. Taschenbuch Elektrotechnik. 1559-1562 Deutsche Norm. Kleinheubacher Berichte. Measuring Process and Storage Tank Level with Radar Technology. Jha. Karl Walter. Band 40 (1997) Pehl. Technische Füllstandsmessung und Grenzstandskontrolle. pp. et al. Merkblatt für die Unfallverhütung. April 1986. April 1999 Von Hippel. Fernandes. Short range devices. Taschenbuch Elektrotechnik. Berlin.3] [Philippow. Erwin.). Universität Siegen.2] [DIN0848-2] [EN300440] [Hippel] [Meinke] [Musch] [Pehl. Hüthig Verlag. 1984 Pehl. 1978 Philippow. Stuttgart. Carlos. Mikrowellentechnik. Norwood. 1995 Meinke. Löcherer.2] [Philippow. Erich.82 “Sicherheitsregeln für Arbeitsplätze mit Gefährdung durch elektromagnetische Felder”. Fachbereich Hochfrequenztechnik. Lange. Artech House. 1991 Draft EN 300 440: Electromagnetic compatibility and radio spectrum matters (ERM). pp. Volume 3. Artech House. Entwurf Januar 1991. Thomas. Detlef. Springer-Verlag.). Ehningen. Eugen (Hrsg. Verlag Technik. Hüthig Verlag. Heidelberg. Band 3: Bauelemente und Bausteine der Informationstechnik. Band 2: Antennen und aktive Bauteile. Berlin. Expert-Verlag. Detlef: Low Power FMCW Radar System for Level Gaging. Band 1: Wellenleitungen und Leitungsbausteine. Klaus (Hrsg. Burkhard: Erzeugung einer hochlinearen analogen Frequenzrampe mit Hilfe von Phasenregelkreisen. 1998 [Bonfig] [Brumbi. Berlin. 256-260. Band 4: Systeme der Informationstechnik. 2000 IEEE MTT-S Symposium Digest. Carlos.). Heidelberg. Verlag Technik. Fassung 8. Rama Kant. Solid Dielectric Horn Antennas. 1984 Philippow.

Verein Deutscher Ingenieure. Cleveland. Stolle. 1999. CRC Press. 25 th EuMC Conference Proceedings (1995). 1974. Verband Deutscher Elektrotechniker (VDI/VDE). 10. New York. SpringerVerlag. Funk. 1993. Christoph: Halbleiter-Schaltungstechnik. Stolle. Skolnik.2] [Tietze] [VDI3519] [Voges. Auflage. Voges. Voges.: Radar Handbook. New York. Schenk. Berlin. CRC Press. 55th Edition.C 3. 1989. Instrumentation and Sensors Handbook. John G. Merrill I.1] [Voges. Berlin. Hüthig Verlag. Cleveland. Auswertemethoden zur Präzisionsentfernungsmessung mit FMCW-Systemen und deren Anwendung im Mikrowellenbereich. Radar-Füllstandsmesssysteme [Schiek] [Skolnik] [Stolle. Edgar. 1991. Heidelberg. Heidelberg.und Radartechnik. Burkhard. Handbook of Chemistry and Physics. Hüthig Verlag. Tietze.1] Schiek. Beuth-Verlag. Band 1: Bauelemente und Schaltungen. Reinhard. Robert C. Teil 1 und Teil 2. VDI/VDE 3519. Webster. McGraw-Hill Professional Publishing. Weast.2] [Weast] [Webster] 64 Radar handbook . Band 2: Leistungsröhren. Schiek. Heidelberg. Füllstandmessung von Flüssigkeiten und Feststoffen. Heidelberg. Edgar. Hochfrequenztechnik. SpringerVerlag. (Editor). (Editor). Burkhard: Grundlagen der Hochfrequenz-Messtechnik. 1991. A General Approach to the Estimation of Measurement Errors in Microwave Range Finding. Ulrich. 2002. 1999. Measurement. Technisches Messen 62 (1995) 2. Antennen und Funkübertragung. New York. Heuermann. Hochfrequenztechnik. [Stolle. Reinhard. Holger. Burkhard. Schiek. Berlin.